Validation of a Predictive Model for Opioid Use After Prolapse Surgery

Objectives: Previously, a predictive model was created to more accurately determine the total mean morphine equivalents (MME) that a patient should require following surgery for pelvic organ prolapse (POP). The predictive model utilizes patient demographic information, as well as patient expectations and experience with pain, to determine the total MME a patient should require in the postoperative period. The objective of this study was to validate the predictive model using a randomized controlled trial.

Methods: This was a single institution randomized controlled trial of women undergoing surgery for pelvic organ prolapse. Patients were randomized to an opioid prescription based on the predictive model or a standard opioid prescription. Preoperatively, patients received acetaminophen, celecoxib, and gabapentin. Patients in both groups received prescriptions for scheduled non-opioid medications including acetaminophen and either ketorolac or ibuprofen. The primary outcome was mean morphine equivalents (MME) remaining after 2 weeks. Secondary outcomes included MME prescribed, MME used and patient satisfaction. A priori it was determined that 51 patients in each arm were required to see a 20% difference in MME prescribed.

Results: Fifty-one patients were randomized to the predictive model group and 51 were randomized to the standard prescription group. Patient and surgical demographics were similar, except for the predictive model group expecting to have more postoperative pain and having a higher percentage of patients with anxiety or depression (Table 1). The median total MME remaining was lower in the predictive model group (10 vs 52.5, p< 0.001). The median MME prescribed at discharge on postoperative day 0 (15 vs 75, p< 0.001) and median total MME prescribed during the entire 2-week postoperative period (20 vs 75, p< 0.001) were also lower in the predictive model group. There was no difference in the total MME used between the groups or in median patient satisfaction with their pain control. There were more patients in the predictive model group who called their surgeon for additional opioids during the postoperative period (19.6% vs 5.9%, p=0.04).

Conclusions: This model provides an effective method to predict the MME a patient will need during the postoperative period after undergoing surgery for POP while maintaining high patient satisfaction with postoperative pain control.